Google to Sign EU AI Code Ahead of August Rules

Google Signs EU AI Code of Practice

In a decisive move ahead of August 2, Google has confirmed it will sign the European Union’s voluntary general-purpose AI code of practice. This framework is designed to help developers align with the bloc’s AI Act by establishing clear processes for documentation, risk management, and data governance. Meta, by contrast, recently declined to join, labeling the legislation “overreach.”

Key Guidelines in the Code

• Provide updated documentation on AI tools and their performance
• Avoid training models on pirated or unauthorized content
• Honor content-owner requests to exclude their works from datasets
• Implement robust risk- and quality-management systems

Google’s Reservations

Despite endorsing the final code, Google’s Kent Walker warned that certain provisions could slow Europe’s AI development. He cited potential conflicts with EU copyright rules, lengthy approval processes, and risks to trade secrets, which could hamper competitiveness and discourage innovation across the region.

Impact on AI Providers

Under the AI Act, major players like Anthropic, Google, Meta, and OpenAI—along with other providers of general-purpose models deemed to have “systemic risk”—must comply within two years. The Act bans unacceptable uses (such as social scoring) and designates high-risk applications (including biometrics and employment tools), all requiring registration and governance measures.

Preparing for Compliance

Organizations should begin by mapping current AI systems against the code’s requirements, identifying data sources, and setting up audit trails. Adopting a risk-based approach—where governance scales with model complexity—helps maintain agility. QuarkyByte’s analytical framework offers tailored workshops, gap assessments, and implementation roadmaps so teams stay on track and competitive.